Customer

The section that everything else depends on and that is most often filled in from imagination. The test for this file is simple: how many of these claims came from a conversation with a real person, and how many came from you thinking hard? Both are allowed. Only one of them may be tagged source-of-truth.

---

The customer, as a person

[TBD]

Not a segment. A person, in a situation.

✗ "Urban millennials interested in wellness."

✓ "A 32-year-old consultant in a large city who works late, has been told by a doctor to fix her gut health, has tried three supplements, and does not have a routine she can keep."

If you have more than one customer type, write each one separately and mark which is primary. The most expensive early mistake is trying to serve three customers with one message.

If you sell B2B, name the three roles

Deals stall because a founder pitched to one person and needed three.

The job to be done

When [context], [customer] needs to [job] so they can [outcome].

[TBD]

Fill in that sentence literally. It is harder than it looks and the difficulty is diagnostic — if you cannot complete it, you do not yet know what you are selling.

What they do today instead

[TBD]

Every customer already solves this problem somehow. They call someone. They use a spreadsheet. They put up with it. They pay a competitor. They ask their cousin.

Include "doing nothing" as a row. It is usually your largest competitor and it never gets named in a competitive analysis.

How often, and how much it hurts

A problem that hurts badly but once a year is a hard business. A mild annoyance that recurs daily can be a great one. Know which you have — it determines your entire pricing and retention model.

Willingness to pay

Stated willingness to pay is close to worthless. "I would definitely buy that" is what people say to be nice to a founder standing in front of them. The only evidence that counts is money that has moved, or a signed commitment to move it.

Interview evidence

Conversations to date: [TBD] · Target: at least 10 before building anything expensive

Use worksheets/customer-interview.md for the guide and the note format. Record the calls where you can — a transcript in _inbox/ is worth ten remembered summaries, and it is what lets the AI find the pattern you missed.

The top five pains, ranked by how often they came up

Verbatim quotes, not paraphrases. The customer's own words are what you will use in your website copy, your ads and your deck — and they are always better than the words you would have written.

What surprised us

[TBD]

The most valuable output of customer research is the thing you did not expect. If ten interviews produced no surprises, you were asking leading questions.

Ideal Customer Profile (ICP)

The targeting filter. This is what turns into a prospect list in 05-gtm.

Estimated count of businesses/people matching this ICP in your beachhead: [TBD] → feeds 03-market

The disqualifier row matters more than founders expect. Knowing who you will not serve is what makes the pitch sharp and stops you burning three months on a customer who was never going to convert.

Objections heard in the field

Every "no" is data. Log it here, not just in your memory.

The same objection appearing five times is your positioning problem, discovered for free. Feed this into 05-gtm collateral and 09-brand messaging.

After the sale

Every customer is a line in a sheet: date, what they bought, satisfaction, what they said, whether they came back. Not a fact somebody on the team half-remembers. This costs nothing to start on day one and is impossible to reconstruct on day four hundred.